Dynamic vehicle routing problem with cooperative strategy in disaster relief
Résumé
Many studies have focused on static vehicle routing problems (VRP) in which all information is known in advance. However, in recent years, the growth of technology has brought about a new range of problems, referred to as dynamic vehicle routing problems. In these problems, a part of the orders are received in advance before departure of vehicles from depots, but some new orders will come in after the vehicle’s departure. In both situations utilising multiple vehicles and cooperative strategy can decrease costs. Although cooperative strategy has not received considerable attention in the literature, it could be a possibility in practice which can help reduce costs. Multiple vehicles in this strategy are allowed to travel and they can transfer goods between one another (from the main vehicle) when they meet in demand points so as to better satisfy the late demands. The benefits of such optimisations, which consider dynamic orders, are evaluated in case of emergencies far from the expected zones and the distribution centres. A mixed integer nonlinear mathematical model is proposed for multi-vehicle routing problem considering product transshipment between vehicles in dynamic situations. The objective function is to minimise the total cost of transportation as well as the number of lost sales. A robust genetic algorithm is then developed to deal with the complexity of the problem. The experimental results show that cooperative strategy is an attractive possibility to reduce unsatisfied late received demands and costs.